Genetic variants in Chinese patients with sporadic dilated cardiomyopathy: a cross-sectional study.

Background: Multiple genes have been associated with familial dilated cardiomyopathy (DCM). However, the role of genetic factors in sporadic DCM (SDCM) remains unclear. Therefore, we studied the genetic variations in Chinese patients with SDCM.
Methods: Sixty-six unrelated Chinese patients (mean age 49.1±17.0 years; 71% male) diagnosed with SDCM were enrolled. The clinical history and genomic DNA of the cohort were collected and examined. The exons of 24 genes closely associated with familial DCM (ABCC9, ACTC1, ACTN2, DES, LAMA4, LDB3, LMNA, MYBPC3, MYH6, MYH7, MYPN, PLN, PSEN1, PSEN2, RBM20, SCN5A, SGCD, TAZ, TCAP, TMPO, TNNI3, TNNT2, TPM1, and VCL) were sequenced using targeted next-generation sequencing method. All called nonsynonymous variants and their occurrence frequencies were compared against population data from public databases. And the nonsynonymous variants were also evaluated for pathogenicity by PolyPhen 2 (PP2) and Sorts Intolerant From Tolerant (SIFT) algorithms.
Results: Eighty-five nonsynonymous variants were detected in 17 genes. The variants and their occurrence frequencies in the patients were compared against population data from the 1000 Genomes and NHLBI (National Heart, Lung, and Blood Institute) Go Exome Sequencing Project. Forty-nine nonsynonymous variants had occurrence frequencies that were significantly higher in the study patients than in the general population, indicating that they have the potential to increase the risk of DCM. The risk variants were distributed in 40 (61%) patients, among whom 25 carried a single variant, while the remaining patients carried multiple (2 to 4) variants. Risk variants occurred more frequently in MYBPC3 (14% of the patients), SCN5A (14%), MYH7 (12%), MYPN (9%), and LDB3 (8%), as verified by Poisson distribution analysis, which were considered "the five risky genes". Conclusions: We found that genetic variants with potential risk for DCM were commonly present in SDCM patients, indicating that genetic factors contribute to the pathogenesis, and (probably) the onset, of DCM in these patients. 2022 Annals of Translational Medicine. All rights reserved.

Introduction

Dilated cardiomyopathy (DCM) is characterized by left ventricular (LV) dilatation and impaired systolic function, and is a leading cause of heart failure (1). Idiopathic DCM is defined when no other discernible causes, such as ischemia, valvular disease, and myocarditis, are present. Epidemiological studies have shown that a proportion of idiopathic DCM cases are familial (2). Familial DCM (FDCM) is primarily transmitted through an autosomal dominant pattern, while other patterns, including X-chromosomal, autosomal recessive, and mitochondrial transmission, are much less common or rare. To date, an increasing number of individual genes have been associated with inheritance in familial DCM cases (1,2).

However, more idiopathic DCM cases are sporadic, and the etiology of sporadic DCM (SDCM) remains largely unknown. SDCM patients are diagnosed with DCM manifestation but no family history, which are common in clinical practice. So far, compared to the information obtained from a large number of studies on FDCM cases, much less information is available regarding the role of genetic factors, such as rare genetic variations, on the pathogenesis of SDCM. Previously, a few studies on mutations of several individual genes implicated the role of genetic factors in the development of SDCM (3-6). Several cohorts, mainly containing Caucasian patients with FDCM or a combination of FDCM and SDCM, were also investigated for mutations in a relatively small number of selected candidate genes, including MYH7, TNNT2, SCN5A, CSRP3, and LBD (7-9).The consideration of genetic variants is important when evaluating the pathogenicity of a genomic variant. The genetic variants were identified using the genome analysis toolkit and the genetic databases. A systematic assessment of correlation between more candidate genes and SDCM cases would help to determine whether SDCM has genetic influences, and if so, whether SDCM shares the same or similar sets of genetic risk factors that are associated with FDCM, and hence, whether these two forms of DCM share similar or major pathogenetic paths. Li et al. reported some genetic variants in 24 SDCM patients recently, while the sample size is small (10). We aimed to expand the subjective scale and study the characteristics of genetic variants in 66 unrelated Chinese patients with SDCM using a next-generation sequencing technique targeting on 24 genes that have been previously reported to be associated with DCM primarily based on familial cases (11,12). We present the following article in accordance with the STROBE reporting checklist (available at https://atm.amegroups.com/article/view/10.21037/atm-21-6774/rc).

Methods

Subjects

All procedures performed in this study involving human participants were in accordance with the Declaration of Helsinki (as revised in 2013). This cross-sectional clinical investigation was conducted in compliance with the guidelines for genetic research in the protocol approved by the Ethics Committees of Zhongshan Hospital (No. 2006-87). All participants signed a written informed consent. All participants in this study were unrelated individuals who were Han Chinese living in eastern China. Patients diagnosed with idiopathic DCM were hospitalized and recruited into the study from the Cardiology Department of Zhongshan Hospital. The inclusion criteria for the SDCM cohort were consistent with the guidelines described by the AHA Scientific Statement and position statement of the ESC working group (13,14). The exclusion criteria included peripartum cardiomyopathy and secondary dilated cardiomyopathies caused by ischaemic heart diseases, hypertension, valvular diseases, endocrine disorders (such as diabetic cardiomyopathy and hyperthyroid cardiomyopathy), inflammation, toxicity, and stress. All patients in this study reported no family history of DCM based on their recollection. Further clinical and genetic screenings were not conducted.

Clinical data collection

Clinical information available from the study subjects included date of birth, gender, vital status, clinical diagnosis, age at diagnosis (age at genetic testing was recorded if the age at diagnosis was not available), family history of DCM and other cardiovascular or muscular diseases, cardiovascular history (including myocardial infarction, hypertension, myocarditis, and drug/toxin exposure), and cardiac structure and function [including maximal LV wall thickness, LV ejection fraction (LVEF), and LV dimensions].

Genomic DNA (deoxyribonucleic acid) sequencing and data analysis

Genomic DNA was isolated from the peripheral blood of the participants. From each sample, 5 µg of genomic DNA was dissolved in 50 µL of water and fragmented to a size of 100–300 bp, as judged by agarose gel electrophoresis. The patients’ DNA samples were screened using a next-generation sequencing technique for point mutation variants in exons of the 24 genes that were reported to be associated with DCM primarily based on FDCM cases. The following genes were included: ABCC9 (ATP-sensitive potassium channel regulatory subunit SUR2A), ACTC1 (cardiac actin), ACTN2 (α-actinin-2), DES (desmin), LAMA4 (laminin a-4), LDB3 (cypher), LMNA (lamin A/C), MYBPC3 (myosin-binding protein C), MYH6 (α-myosin heavy chain), MYH7 (β-myosin heavy chain), MYPN (myopalladin), PLN (phospholamban), PSEN1 (presenilin 1), PSEN2 (presenilin 2), RBM20 (RNA binding protein 20), SCN5A (cardiac sodium channel), SGCD (δ-sarcoglycan), TAZ (tafazzin), TCAP (telethonin), TMPO (thymopoietin), TNNI3 (cardiac troponin I), TNNT2 (cardiac troponin T), TPM1 (tropomyosin α-1 chain), and VCL (metavinculin). Next-generation sequencing was performed at the sequencing platform of the Shanghai Institute of Cardiovascular Diseases.

Variants were called by aligning the raw sequence data to the human GRCh37 reference genome with manual verification. Novel variants were defined as those variants that were not found in the SNP Database (dbSNP) build 137 of the National Center for Biotechnology Information, and that had not been reported at the completion of the study. All called nonsynonymous variants of the genes in the DCM cohort and their occurrence frequencies were compared against population data from public databases, including the 1000 Genomes Project (http://www.1000genomes.org), and NHLBI (National Heart, Lung, and Blood Institute) Go Exome Sequencing Project (http://evs.gs.washington.edu/EVS/). The sequencing data of the 197 Chinese adults deposited in the 1000 Genomes Project database were used as the Chinese (ethnic-matched) control group in this study. Sequence and occurrence frequency data from other populations in this database, and those from African American and European American populations in the NHLBI Go Exome Sequencing Project database were combined together and used as the control group of the non-Chinese populations.

The relationship between the nonsynonymous variants and DCM was evaluated by comparing the occurrence frequencies of the DCM cohort and the control populations. Details of the comparison are described in the Results section below. The nonsynonymous variants were also evaluated for pathogenicity by PolyPhen 2 (PP2) and Sorts Intolerant From Tolerant (SIFT) algorithms. PP2 scores were obtained using HumVar model software (http://genetics.bwh.harvard.edu/pph2/index.shtml). PP2 scores ranged from 0 to 1, with three levels of pathogenic potential, i.e., probably damaging, possibly damaging, and benign). SIFT scores were also obtained using a software tool (http://siftdna.org/www/Extended_SIFT_chr_coords_submit.html). The SIFT scores range from 0 to 1, and were divided into damaging or tolerated levels with 0.05 as the threshold value.

Statistical analysis

Normally distributed continuous data were presented as mean ± standard deviation (SD) and compared using the Student t-test. However, if the normality test (Shapiro-Wilk) and/or equal variance test failed, the Mann-Whitney rank-sum test was used to compare the continuous data. Categorical variables were presented as frequencies and analyzed by Fisher’s exact test using the Simple Interactive Statistical Analysis web-based software (SISA Binomial, Southampton, UK). In all statistical tests, P<0.05 was used to determine whether differences were statistically significant. To evaluate the occurrence frequencies of genetic variants, a model of simple Poisson process was used. We assumed that the number of variants occurring in a given gene approximately follows the Poisson distribution, p(k) = e−λ λ/k!. P<0.05 indicated a significantly high occurrence frequency. Poisson distribution and P values were computed using a web-based tool (http://www.vassarstats.net/poissonfit.html#down). The same method was also used to evaluate the significance of the prevalence of genetic variants.

Results

Study cohort

The SDCM cohort included 66 patients (47 males). The average age at diagnosis of patients in this cohort was 49.1±17.0 years. All enrolled patients had symptoms of heart failure at the time of enrollment, with a majority of them in the New York Heart Association’s (NYHA) classes III and IV. Echocardiographic findings showed a mean LV end-diastolic diameter (LVEDD) of 69±9 mm, a mean LV end-systolic diameter (LVESD) of 57±10 mm, and a mean LVEF of 34%±10%.

Nonsynonymous variants of DCM-associated genes in the SDCM cohort

A total of 85 nonsynonymous variants, mostly single-nucleotide mutations, were detected by sequencing in 17 out of the 24 DCM-related genes in the DCM cohort (listed in ). In this cohort, nonsynonymous variants were absent in seven DCM-related genes. Fifty-five (65%) of the called nonsynonymous variants were either registered in the databases we searched or were reported previously (we denoted these variants as ‘known variants’). The remaining 30 (35%) of the called nonsynonymous variants have not yet been recorded anywhere else, and were therefore denoted herein as ‘novel variants’. All novel variants were rare frequency mutations (i.e., allele frequency <0.5% in the combined control populations). Meanwhile, among the 55 known variants, 19 (35%) had higher allele frequencies (>0.5%) in the combined control populations.

Table 1

Nonsynonymous variants of the DCM-associated genes found in the sporadic DCM cohort and their occurrence frequency compared with control populations

Gene	dbSNP ID	Change in AA	Number of DCM patients	DCM (D) allele counts	Chinese (C) allele counts	All Chinese (AC) allele counts	Non-Chinese (NC) allele counts	All reference population (A) allele counts	P (D-C)	P (D-NC)	P (C-NC)	P (D-A)	P (AC-NC)	Pattern	Disease reported
!ABCC9	rs149319186	K976I	1	1/132	2/394	3/526	0/14,796	2/15,190	1.0000	0.0088	0.0007		0.0000*	2
†ABCC9		R1197C	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†ACTN2		K96R	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†ACTN2		M316T	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!ACTN2	rs80257412	D475N	6	6/132	35/394	41/526	20/14,796	55/15,190	0.1334	0.0000	0.0000		0.0000*	2
†DES		R78L	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†LAMA4		A41V	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†LAMA4		C91S	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!LAMA4	rs71543223	A283D	54	108/132	343/394	451/526	1,444/1,790	1,787/2,184	0.1507	0.8195	0.0024		0.0083*	2
!LAMA4	rs1050348	Y498H	65	107/132	309/394	416/526	9,713/1,4796	10,022/15,190	0.6209	0.0001	0.0000		0.0000*	2
!LAMA4	rs2032567	G1117S	66	116/132	338/394	454/526	11,172/14,796	11,510/15,190	0.6609	0.0007	0.0000		0.0000*	2
!LAMA4	rs1050349	P1119R	34	41/132	144/394	185/526	3,252/14,796	3,396/15,190	0.2924	0.0151	0.0000		0.0000*	2
!LAMA4	rs70940811	V1315I	1	1/132	6/394	7/526	4/14,796	10/15,190	0.6860	0.0434	0.0000		0.0000*	2
†LAMA4		G1356R	2	2/132	0/394	2/526	0/14,796	0/15,190	0.0626	0.0001	1.0000	0.0001*		1
!LAMA4	rs201094782	Y1391H	1	1/132	1/394	2/526	0/14,796	1/15,190	0.4393	0.0000	0.0259		0.0012*	2
!LAMA4	rs3734292	V1815I	11	12/132	31/394	43/526	15/14,796	46/15,190	0.7136	0.0000	0.0000		0.0000*	2
†LDB3		T28K	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!LDB3	rs3740343	V55I	7	7/132	38/394	45/526	23/14,796	61/15,190	0.1053	0.0000	0.0000		0.0000*	2	LVNC (15)
†LDB3		E139K	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!LDB3	rs45521338	R218C	1	1/132	2/394	3/526	1/14,796	3/15,190	1.0000	0.0176	0.0020		0.0002*	2
†LDB3		S330P	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!LDB3	rs113817827	V426I	1	1/132	2/394	3/526	2/14,796	4/15,190	1.0000	0.0263	0.0001		0.0004*	2
†LDB3		M456R	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†LDB3	rs145983824	P498L	1	1/132	0/394	1/526	3/14,796	3/15,190	0.2510	0.0349	1.0000	0.0340*		1
†LMNA		R220H	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!MYBPC3	rs3729989	S236G	2	2/132	13/394	15/526	1,475/14,348	1,488/14,742	0.3769	0.0001	0.0000		0.0000*	2	HCM (16)
†MYBPC3		E334K	4	5/132	0/394	5/526	0/14,796	0/15,190	0.0009	0.0000	1.0000	0.0000*		1	HCM (17)
†MYBPC3		R409G	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†MYBPC3		G416S	1	1/132	0/394	1/526	3/14,466	3/14,860	0.2510	0.0357	1.0000	0.0348*		1	HCM (18)
†MYBPC3		P459fs	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1	HCM (19)
†MYBPC3		R835L	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†MYBPC3		R895H	1	1/132	0/394	1/526	4/14,078	4/14,472	0.2510	0.0456	0.1045	0.0444*		1
!MYH6	rs28711516	G56R	1	1/132	10/394	11/526	1,155/14,796	1,165/15,190	0.3060	0.0005	0.0001		0.0000*	2
!MYH6	rs365990	V1101A	21	21/132	68/394	89/526	6,644/14,796	6,712/15,190	0.7894	0.0000	0.0000		0.0000*	2
MYH6	rs28730771	A1130T	6	6/132	27/394	33/526	168/1,790	195/2,184	0.2990	0.0600	0.1188	0.1087		3
!MYH6	rs34935550	E1295Q	2	2/132	10/394	12/526	43/14,796	53/15,190	0.7389	0.0600	0.0000		0.0000*	2
MYH6	rs45574136	Q1593L	2	2/132	11/394	13/526	327/14,796	338/15,190	0.5328	1.0000	0.3877	1.0000		3
MYH6	rs61742476	V1613A	1	1/132	11/394	12/526	248/14,796	259/15,190	0.3106	0.7288	0.1092	0.7299		3
†MYH6		K1860R	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†MYH7	rs121913653	T441M	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1	Distal myopathy (20)
†MYH7	rs121913625	R453C	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1	HCM (21)
†MYH7		I736T	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1	HCM (22)
†MYH7	rs121913628	E924K	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1	HCM (18)
†MYH7		LS1139LD	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†MYH7		R1250W	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1	LVNC (23)
†MYH7		G1520R	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†MYH7		R1897H	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†MYPN		R482K	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
!MYPN	rs10823148	F628L	23	27/132	67/394	94/526	5,819/14,796	5,886/15,190	0.3617	0.0000	0.0000		0.0000*	2
!MYPN	rs10997975	S691N	25	31/132	86/394	117/526	5,801/14,796	5,887/15,190	0.3617	0.0002	0.0000		0.0000*	2
!MYPN	rs7916821	S707N	22	27/132	67/394	94/526	5,788/14,796	5,855/15,190	0.3617	0.0000	0.0000		0.0000*	2
!MYPN	rs3814182	S803R	36	46/132	134/394	180/526	7,714/14,796	7,848/15,190	0.9156	0.0001	0.0000		0.0000*	2
!MYPN	rs181848049	G847V	1	1/132	2/394	3/526	0/14,796	2/15,190	1.0000	0.0088	0.0007		0.0000*	2
†MYPN	rs151282801	R1042C	1	1/132	0/394	1/526	3/14,796	3/15,190	0.2510	0.0349	1.0000	0.0340*		1
!MYPN	rs7079481	P1135T	29	33/132	82/394	115/526	5,983/14,796	6,065/15,190	0.3312	0.0002	0.0000		0.0000*	2
†MYPN	rs138313730	L1161I	3	3/132	0/394	3/526	12/14,796	12/15,190	0.0155	0.0003	1.0000	0.0003*		1
†MYPN	rs199585352	S1296T	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†RBM20		G40W	1	1/132	0/394	1/526	0/6,356	0/6,750	0.2510	0.0204	1.0000	0.0192*		1
†RBM20		P48delinsPP	1	1/132	0/394	1/526	0/6,356	0/6,750	0.2510	0.0204	1.0000	0.0192*		1
!RBM20	rs143785916	R641Q	1	1/132	3/394	4/526	2/6,356	5/6,750	1.0000	0.0598	0.0018		0.0000*	2
RBM20	rs138926584	R673Q	1	1/132	0/394	1/526	7/6,356	7/6,750	0.2510	0.1517	1.0000	0.1436		3
RBM20	rs1417635	W768S	66	132/132	394/394	526/526	6,125/6,127	6,519/6,521	1.0000	1.0000	1.0000	1.0000		3
†RBM20		Q856X	1	1/132	0/394	1/526	0/6,356	0/6,750	0.2510	0.0204	1.0000	0.0192*		1
!RBM20	rs188054898	R1057Q	4	4/132	11/394	15/526	5/6,356	16/6,750	1.0000	0.0001	0.0000		0.0000*	2
†RBM20		R1182H	1	1/132	0/394	1/526	0/6,356	0/6,750	0.2510	0.0204	1.0000	0.0192*		1
!RBM20	rs942077	E1223Q	59	101/132	315/394	416/526	4719/6,356	5,034/6,750	0.3902	0.6154	0.0122		0.0142*	2
†SCN5A	rs199473071	R225Q	3	3/132	0/394	3/526	0/14,118	0/14,512	0.0155	0.0000	1.0000	0.0000*		1
†SCN5A	rs199473561	A226V	1	1/132	0/394	1/526	0/14,118	0/14,512	0.2510	0.0093	1.0000	0.0090*		1	BS (18)
!SCN5A	rs1805124	H558R	16	16/132	38/394	54/526	3,520/14,410	3,558/14,804	0.4110	0.0007	0.0000		0.0000*	2	AF, LQTS (24)
†SCN5A	rs45600438	R568C	1	1/132	0/394	1/526	0/14,264	0/14,658	0.2510	0.0089	1.0000	0.0086*		1
†SCN5A		R659W	1	1/132	0/394	1/526	0/14,784	0/15,178	0.2510	0.0093	1.0000	0.0090*		1	LQTS (25)
!SCN5A	rs1805125	P1090L	3	3/132	9/394	12/526	4/14,236	13/14,630	1.0000	0.0000	0.0000		0.0000*	2	LQTS (25)
†SCN5A	rs41310765	A1180V	1	1/132	1/1,314	2/1,446	0/14,794	1/15,188	0.4393	0.0088	0.0816	0.0162*		1	AF, DCM (26)
!SCN5A	rs41261344	R1193Q	8	10/132	20/394	30/526	17/14,796	37/15,190	0.2832	0.0000	0.0000		0.0000*	2	BS, LQT, CCD, DCM (27)
†SCN5A	rs199473251	I1448N	1	1/132	0/394	1/526	0/14,122	0/14,516	0.2510	0.0092	1.0000	0.0089*		1
†SCN5A		P1619T	1	1/132	0/394	1/526	0/14,796	0/15,190	0.2510	0.0088	1.0000	0.0086*		1
†SGCD		C34G	1	1/132	0/394	1/526	0/13,987	0/14,381	0.2510	0.0094	1.0000	0.0091*		1
†SGCD		P253fs	1	1/132	0/394	1/526	0/13,890	0/14,284	0.2510	0.0094	1.0000	0.0092*		1
†SGCD		Q283R	1	1/132	0/394	1/526	0/14,022	0/14,416	0.2510	0.0093	1.0000	0.0091*		1
†TAZ		H111N	1	1/132	0/394	1/526	0/12,353	0/12,747	0.2510	0.0106	1.0000	0.0103*		1
†TNNT2		R151Q	2	2/132	0/394	2/526	0/14,310	0/14,704	0.0626	0.0001	1.0000	0.0001*		1
TNNT2	rs3730238	K260R	4	4/132	14/394	18/526	889/14,796	903/15,190	1.0000	0.1946	0.0400		0.1452	3	HCM (28)
†TPM1		D34E	1	1/132	0/394	1/526	0/14,308	0/14,702	0.2510	0.0091	1.0000	0.0089*		1
!VCL	rs144683137	M209L	1	1/132	1/394	2/526	0/14,796	1/15,190	0.4393	0.0088	0.0259		0.0012*	2
!VCL	rs201528612	P398S	1	1/132	1/394	2/526	0/14,796	1/15,190	0.4393	0.0088	0.0259		0.0012*	2

DCM patients: the patients with sporadic DCM enrolled in this study; Chinese: the 197 Chinese adults in the 1000 Genomes Project (phase 1), for A1180V of SCN5A samples from additional 460 unrelated healthy Chinese (29) were included; all Chinese: DCM patients + Chinese; non-Chinese: a combination of populations excluding the Chinese population in the 1000 Genomes Project (phase 1) and American populations in NHLBI Go Exome Sequencing Project; all control populations: Chinese + non-Chinese, referred as the general control population in the text. Patterns of the grouped variants: †, pattern 1, having potential risk of sporadic DCM: when P(C-NC) ≥0.05; a variant was judged to be in this pattern group if P(D-A) <0.05. !, pattern 2, being Chinese-specific without risk of sporadic DCM: when P(C-NC) <0.05 and P(D-C) ≥0.05; groups D and C were pooled together to form group AC, and a variant was judged to exhibit this pattern if P(AC-NC) <0.05. Pattern 3, being shared globally without risk of sporadic DCM: a variant that falls neither in pattern 1 nor in pattern 2. *, P<0.05. AA, amino acid; DCM, dilated cardiomyopathy; LVNC, left ventricular noncompaction; HCM, hypertrophic cardiomyopathy; AF, atrial fibrillation; LQTS, long QT syndrome; BS, Brugada syndrome.

Occurrence frequencies of the nonsynonymous variants

To determine the possible relationship between the nonsynonymous variants and DCM in SDCM patients, we first compared the occurrence frequency (allele frequency) of the variants identified in the DCM cohort of this study with those of the control populations. We used two control populations, i.e., Chinese and non-Chinese controls, as described in the Methods section above. In addition, we also combined the Chinese and non-Chinese controls to form the general control population. The comparison indicated that the variants could be divided into three groups with statistically separable patterns.

The allele frequencies of variants in the first pattern group in the SDCM cohort were significantly higher than those in the general control population, but were statistically similar between the Chinese and non-Chinese control populations. These statistical features suggest that the variants falling in this pattern group exhibit a potential risk for SDCM (herein, we denoted these variants as ‘risk variants’). There were 49 variants (58% of the total nonsynonymous variants called) in this risky group. Most risk variants were not present in the control populations, except for six variants that had very low allele frequencies in the control populations. All risk variants were heterozygous mutations, except for a patient who carried a homozygous E334K mutation. It is noteworthy that all 30 novel variants were risk variants.

The allele frequencies of the variants in the second pattern group were statistically different between the Chinese and non-Chinese populations, but were statistically similar between the SDCM group and the Chinese control group. These results suggest that variants with these statistical features (30 variants, 35% of the total variants called) exhibit specific occurrence in the Chinese population, but are not likely to indicate an increased risk of SDCM. We found that the allele frequencies of variants that did not fall in the previous two pattern groups in the SDCM cohort were statistically similar to those in the general control population, indicating that these six variants (7% of the total variants called) were shared globally and do not exhibit a specific risk of SDCM. We denoted variants in the second and third groups as ‘low-risk variants’. displays the statistical data that are summarized here.

The risk variants were distributed in 16 DCM-associated genes, which accounted for 94% of the genes on which nonsynonymous variants were found, and 67% of the total number of genes in our study. They were distributed in 40 (61%) patients, among whom 25 carried a single variant, 11 carried two variants, two carried three variants, and one carried four variants. As shown in , the prevalence of risk variants of each gene ranged from 1 to 9 among the 66 patients. MYBPC3 and SCN5A exhibited the highest prevalence (9/66, 14%). Poisson distribution analysis revealed that MYBPC3, SCN5A, MYH7, MYPN, and LDB3 had significantly higher prevalence (14%, 14%, 12%, 9%, and 8%, respectively) than the other genes. They had a combined prevalence of 31/66 (47%) in the SDCM cohort when multiple risk variants of these genes in a patient were counted as a single occurrence. MYBPC3, SCN5A, MYH7, and LDB3 also housed more risk variants than the other genes, and the Poisson distribution analysis identified that they had a significantly high probability to house risk variants. These results suggested that these genes (which we denoted as ‘risky genes’) were closely related to SDCM.

Table 2

Genes hosting variants with potential risk of sporadic DCM in the DCM cohort being studied

Gene	Prevalencea of variants with risk of sporadic DCM	P valuec	Prevalencea of variants	Number of variants with risk of sporadic DCM	P valued	Number of total variants
MYBPC3	9 (13.6%)b	0.000*	11	6	0.010*	7
SCN5A	9 (13.6%)	0.000*	36	7	0.002*	10
MYH7	8 (12.1%)	0.001*	8	8	0.001*	8
MYPN	6 (9.0%)	0.011*	142	4	0.140	10
LDB3	5 (7.6%)	0.039*	14	5	0.034*	8
RBM20	4 (6.1%)	0.114	135	4	0.140	9
LAMA4	4 (6.1%)	0.114	236	3	0.261	10
SGCD	3 (4.5%)	0.277	3	3	0.261	3
ACTN2	2 (3.0%)	0.546	8	2	0.528	3
TNNT2	2 (3.0%)	0.546	6	1	0.830	2
MYH6	1 (1.5%)	0.840	34	1	0.830	2
ABCC9	1 (1.5%)	0.840	2	1	0.830	7
DES	1 (1.5%)	0.840	1	1	0.830	1
LMNA	1 (1.5%)	0.840	1	1	0.830	1
TPM1	1 (1.5%)	0.840	1	1	0.830	1
TAZ	1 (1.5%)	0.840	1	1	0.830	1
VCL	0 (0%)	1.000	2	0	1.000	2
ACTC1	0 (0%)		0	0		0
PLN	0 (0%)		0	0		0
PSEN1	0 (0%)		0	0		0
PSEN2	0 (0%)		0	0		0
TCAP	0 (0%)		0	0		0
TMPO	0 (0%)		0	0		0
TNNI3	0 (0%)		0	0		0

a, prevalence was designated as the sum of patients carrying a variant (either homozygous or heterozygous) in each gene. When prevalence was calculated, variants co-occurring in a gene were counted separately. Likewise, variants co-occurring in a patient were also counted separately. b, prevalence in this column was also expressed as a percentage of the total patients in the DCM cohort. c,d, P values were computed using Poisson distribution as described in the Methods section. The means of the fitted Poisson distribution (λ) were 0.97 and 0.95, respectively. Only those genes that contained nonsynonymous variants were included in the computation. *, P<0.05.

Pathogenic potential of the nonsynonymous variants

Seventeen known variants, 11 risky variants, and six low-risk variants have been previously found to increase susceptibility to some cardiac diseases, including Brugada syndrome, LV noncompaction, hypertrophic cardiomyopathy, and distal myopathy (). However, none of the variants were directly related to DCM, except for A1180V and R1193Q of SCN5A, which were reported in the DCM cases (26,30).

To obtain further information on the pathogenic potential of the nonsynonymous variants of the DCM-associated genes in the SDCM cohort, we performed an analysis of the nonsynonymous variants using the protein function prediction algorithms, PP2 and SIFT (Table S1). In total, 62 variants could be predicted by both algorithms (73% of the total nonsynonymous variants), which also included one non-sense and two frame-shift variants that could be assumed to cause damaging consequences in the protein function (31). Twenty-six risk variants (53% of the total risk variants) were predicted to be damaging, and eight risk variants (16% of the total risk variants) were predicted to be tolerated. On the other hand, 23 low-risk variants (63% of the total low-risk variants) were predicted to be tolerated, and five low-risk variants (14% of the total low-risk variants) were predicted to be damaging. Fisher’s exact test indicated a significant difference (P=0.000) in the predicted results between the risky and non-risky variant groups. Therefore, the protein function prediction algorithms confirmed that majority of the risk variants exhibit pathogenic potential, whereas the majority of low-risk variants were predicted to be harmless based on the protein structure-function relationship. We noticed that in the risky genes of SDCM, SCN5A housed the most risk variants that were predicted to be damaging.

Comparison of clinical symptoms between the patients with and without risk variants

Given the determination of the risk variants, we further explored whether there were differences in clinical symptoms between the patients with and without risk variants, and compared the symptoms between these two groups of patients (as shown in Tables S2,S3). The results indicated that patients with risk variants did not manifest any symptoms or abnormalities that were significantly different from those seen in the patients without risk variants.

Discussion

This study was specifically designed to screen genes known to be associated with FDCM in SDCM patients. To our knowledge, similar studies have not yet been reported in the literature. In this study, we investigated a cohort of 66 unrelated Chinese patients with diagnosed SDCM. We performed mutational screening of 24 genes known to be associated with FDCM using a next-generation sequencing technique. A major finding of this study was that the at-risk genotypes are common (61%) in SDCM patients. The second major finding was that MYBPC3 and SCN5A were found to be the most prevalent risky genes for SDCM. The five significantly risky genes out of the 24 genes, namely MYBPC3, SCN5A, MYH7, MYPN, and LDB3, had a combined prevalence of 47% in the patient cohort, which accounted for 78% of the prevalence of the total risk variants. In addition, 30 novel variants with a potential to increase the risk of DCM were identified in this study.

Comparison of the prevalence of risk variants with other studies

The prevalence of 61% for total risk variants observed in this study was higher than that reported in previous studies that included both FDCM and SDCM. In a study by Millat et al., a prevalence of 19% was observed in 105 DCM patients (8). Another study by Hershberger et al. was 11.5% in 313 DCM patients, although the study cohorts of these studies contained both FDCM and SDCM patients (32). We consider that these differences are partly due to the fact that our study included more genes than these previous studies. For example, only six genes (MYH7, TNNT2, SCN5A, TCAP, LDB3, and CSRP3) were screened in the 313-case study, which excluded two genes with high prevalence (MYBPC3 and MYPN) found in our study. Given the fact that there are more DCM-associated genes, the actual prevalence of genetic risk variants in SDCM patients would be expected to be even higher. Indeed, a later study from Hershberger et al. including additional genes further expanded the prevalence (up to 27%) of the total variants that are likely cause DCM (7). The second factor contributing to this difference is the criteria used to define risk variants. In this study, we used a statistics-based criterion that compares prevalence; a variant was considered to be risky when its prevalence in the DCM cohort was significantly higher than that in the general population (on the condition that the prevalence of this variant in the Chinese population was similar to that in the general population). However, more complicated criteria, which possibly depend on familial cases to a greater extent, were used previously. For example, in the 313-case study, a variant was considered disease-causing if it caused a change in a conserved amino acid, a frame-shift, premature truncation, a mis-splicing event, and also segregated with the disease in multiple affected individuals or was identified in multiple unrelated probands, or had previously been reported to be associated with DCM (32). We believe that the criteria used in the present study are more appropriate for detection of novel risk variants, especially in sporadic cases where segregation of multiple occurrence is unlikely. In the risk variants identified by this criterion, 51% were predicted to be functionally damaging by both the PP2 and SIFT algorithms (in addition to the frame-shift and non-sense variants), and there were significantly more predicted damaging variants in the risky group than in the non-risky group, with a P value of 0.000. These facts confirm the effectiveness of the criteria used in our study. With the rapid expansion of population genomic databases, the accuracy of identification of risk variants using statistics-based criteria should be rapidly increased.

Although our study reported a high prevalence of risk variants, the major fraction of the prevalence was attributed to the variants hosted in a few genes. The five risky genes with the highest prevalence (MYBPC3, SCN5A, MYH7, MYPN, and LDB3) were distributed in 47% of the total SDCM patients and 78% of patients with risk variants. This distribution overlaps with the five genes with the highest prevalence (LMNA, MYBPC3, MYH7, MYH6, and TNNT2) previously obtained from a mixed cohort mainly containing Caucasian patients with FDCM and SDCM. Determination of the genes with the highest prevalence of risky genes may assist in clinical practice by narrowing down the genes to be screened and saving the cost of diagnostic tests. Although the 24 candidates are not the whole targeted genes of SDCM, the results of the study showed the genetic variants from the candidate genes, suggesting the genetic variants are common in DCM even though the patients with no familial history. Besides the environment factors, the genetic factor plays an important role in the pathogenesis of SDCM. The findings of the research indicate the genetic screening is valuable for DCM patients in clinical practice. In addition, our study used next-generation sequencing technology rather than chain termination to determine the DNA sequence, which is faster and more cost-efficient (29). The accurate sequencing results in an array of multiple genes obtained in our study confirm that this high-throughput sequencing technology facilitates diagnostic classification and can improve risk stratification in affected patients.

It is important to note that in this study, the DCM patients were diagnosed as sporadic cases primarily based on the patients’ recollection of their family history, which is sometimes inaccurate, and was not verified by medical examination. Therefore, the DCM cases in this study can only be considered as “apparently” sporadic.

Are genetic factors a risk for DCM in sporadic cases?

The role of genetic factors in the pathogenesis of SDCM remains largely unknown. Previous population studies mainly from Caucasian population showed that the prevalence of disease-causing mutations of some DCM-associated genes is similar for SDCM and FDCM (7,30). Our study conducted in a different ethnic group, namely Chinese, are therefore helpful to further map out the role of genetic factors in SDCM.

Mutational screening studies, such as the present study, provide clear evidence to show that genetic variation has a relation to SDCM. However, compared to the familial cases, it is more difficult to determine the contribution of genetic factors in sporadic cases that do not involve family history, i.e., cases where the disease is not apparently hereditary. Several hypothetical interpretations could be offered. Firstly, it appears that each individual risky variant only modestly increases the risk, which is not sufficient to result in the onset of the disease by itself. Therefore, those who carry multiple risk variants are likely to develop the disease. Indeed, 14 of the 40 (35%) patients with risk variants in our study were found to carry multiple risk variants. We cannot exclude the possibility of patients with single risk variants concurrently carrying risk variants of DCM-associated genes, either known or unknown at present, that are not included in our study.

In addition to genetic factors, environmental factors may play a significant role in the expression of pathogenic mechanisms. Our previous study on a segregated DCM family with A1180V of SCN5A (which encodes the cardiac sodium channel) demonstrated that the functional phenotype of this genetic variation was specifically aggravated at high heart rates (26). The results of that study suggested that the risk of DCM of this variation increases in those carriers with physical activity and lifestyles that increase the average daily heart rate. Notably, A1180V of SCN5A was found to be a risky variant in the present study.

Is there a specific phenotype that differentiates SDCM patients with risk variants from those without?

A unique result of this study is the report of summarized symptoms and the statistical comparison of the symptoms between patients with and without risk variants (shown in Tables S2,S3). Our data clearly demonstrated that there were no differences in the profile of DCM-related symptoms between these two patient groups. DCM was the only phenotype for individuals who carried risk variants reported in our study. No other clear genotype-phenotype correlations could be concluded from the data. A clear unique phenotype-genotype relationship does not appear to be present for individual genes either. For example, many DCM patients with variants in LMNA and SCN5A were reported to suffer from conduction system diseases (26,33); however, in our study cohort, only one patient had first-degree atrioventricular block, whereas other patients carrying risk variants in LMNA and SCN5A were not diagnosed with any conduction system diseases. Also, genes listed among the highest prevalence encode proteins (myosin-binding protein C, cardiac sodium channel, β-myosin heavy chain, myopalladin, and cypher), which have markedly different functions, implicate the heterogeneity of the phenotype-genotype relation. In addition, we noticed that patients carrying multiple variants did not exhibit aggravated symptoms (see Table S2), indicating an absence of additive effects for risk variants on the symptoms of the disease. Certainly, we do not exclude the possibility that some individual variants may not directly cause DCM. However in these cases, DCM is not primary, which is not the concern of this study. Taken together, our data supports the hypothesis that the presence of a risky genotype in SDCM only imposes an increased risk for the onset of DCM, but does not affect the progress of DCM after the onset of the disease. However, the current study does not contain sufficient information regarding the mechanism through which risk variants lead to the onset of SDCM, which was not the purpose of this study and requires further investigation.

Conclusions

In summary, this study suggests that at-risk genomic variants are a major pathogenic factor of SDCM, and MYBPC3, SCN5A, MYH7, MYPN, and LDB3 are the major genes hosting the at-risk genomic variants. This study also identifies a number of novel variants that are possibly associated with SDCM. These results not only expand the spectrum of DCM genetics, but also provide new information that helps to provide insights into the pathogenesis of SDCM.

The article’s supplementary files as